unwinding-codebase

Coordinate subagents to generate layered codebase documentation from YAML architecture.

9|Updated Jan 15, 2026
One-click install
npx skills add https://github.com/cliftonc/unwind --skill unwinding-codebase
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: unwinding-codebase
Source: https://github.com/cliftonc/unwind/tree/main/skills/unwinding-codebase
Command: npx skills add https://github.com/cliftonc/unwind --skill unwinding-codebase

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Orchestrates layer-by-layer analysis of a codebase by coordinating specialized subagents to generate structured, machine-readable layer docs.

Core Features & Use Cases

  • Coordinated layer analysis: Dispatches subagents to analyze database, domain-model, service, API, messaging, frontend, and tests.
  • Incremental documentation: Produces docs/unwind/layers/[layer]/* with index.md and per-layer section files incrementally.
  • Deterministic workflow orchestration: Defines phases, dependencies, and parallelism to standardize unwind process.

Quick Start

Start the unwind workflow to generate per-layer documentation.

Frequently Asked Questions about unwinding-codebase

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate layer-by-layer codebase documentation automatically?

Generate layer-by-layer codebase documentation by dispatching specialized subagents to analyze database, domain-model, service, API, messaging, frontend, and tests. It produces structured docs/unwind/layers/* directories with index.md and per-layer section files incrementally.

What is codebase unwinding and when do I need it?

Codebase unwinding is a layer-by-layer analysis process that extracts YAML layers, builds dependency graphs, and outputs machine-readable layer documentation. You need it when standardizing architecture documentation for synthesis across complex codebases.

Do I need an architecture.md file before starting layer analysis?

Yes, layer analysis requires docs/unwind/architecture.md as input. The orchestrator extracts YAML layers from this file after unwind:start to coordinate subagents and enforce phase order for incremental documentation writes.

How does subagent orchestration work for codebase analysis?

Subagent orchestration coordinates specialized agents through deterministic workflow phases with defined dependencies and parallelism. It standardizes the unwind process by dispatching agents to analyze specific layers and enforcing incremental writes to docs/unwind/layers/[layer]/*.

What's the best way to structure machine-readable layer documentation?

Structure machine-readable layer documentation using docs/unwind/layers/[layer]/index.md and per-layer section files. This format supports dependency graphs and YAML layer extraction, making outputs suitable for downstream synthesis.

What are the limitations of automated layer analysis for large codebases?

Automated layer analysis operates after unwind:start and requires architecture.md as a prerequisite input. It enforces phase order strictly, meaning incomplete prerequisite phases will block downstream layer documentation generation.